Contribution of sensorimotor beta oscillations during value-based action selection

Abstract

Choosing a course of daily-life actions requires an accurate assessment of the associated risks and potential rewards. We investigated the neural dynamics of this decision process by analyzing the neural electrical signals acquired from electroencephalography (EEG) during a value-based action-selection task. In particular, we determined whether sensorimotor beta oscillations, traditionally studied in the context of motor control, are also involved in value-based decision making for actions. Additionally, we examined the involvement of this beta signal relative to other neural signals such as the ERP components P2 and P3, which have been previously identified in reward processing and value computations. Our results from healthy young adults (N = 31), showed a significant decrease in sensorimotor beta power during a decision phase without any motor response, in addition to an action phase when a response was made. The decision-phase beta signal was preceded by the P2/P3b components, and all of these neural signals reliably dissociated the different reward and risk levels, suggesting the encoding of decision variables. Importantly, while the beta signal during both the action and decision phase predicted behavioral performance (i.e., response time) in the action phase, the preceding P2/P3b had no such predictive association with the behavior. Collectively, these results demonstrate a unique contribution of the motor system in value-based decision making for actions, via the translation of motivational information into a motor signal across time.

Keywords: Decision making; Motor control; Reward; Risk; Sensorimotor beta oscillations.